Signal Processing Theory And Algorithms In The Transmit Beamspace-Based MIMO Radar

Multiple-input multiple-output(MIMO)radar has become the focus of intensive research in recent years.Based on the types of array configuration,it can be devided into two regimes which are named as statistical(or distributed)MIMO radar and coherent(or colocated)MIMO radar.It has been verified in the literature that the former can enhance the ability to combat signal scintillation,and therefore can imporove the target detection as well as parameter estimation performance,while the latter enables improving angular resolution,bettering upper limit on the number of detectable targets,enhancing parameter identifiability,extending the array aperture via virtual sensors,and increasing opportunities of clutter and jammer suppression.We focus our studies on the regime of colocated MIMO radar throughout this thesis.To be specific,we first review the relevant and newest research on MIMO radar which covers the topics such as waveform design,transmit beamforming(i.e.,transmit beamspace design),clutter and jammer suppression,target detection and parameter estimation,representative experiments,and other issues emerged in recent years.Then we compare the spatial performance(i.e.,non-adaptive and adaptive beamforming performance)of phased-array(PA)radar,phased-MIMO radar whose transmit array is separated into overlapped or non-overlapped subarrays,and standard(or conventional)MIMO radar.General signal model has been established for these radar configurations,based on which we derive the corresponding beampattern gain and output signal-to-interference-plus-noise ratio(SINR)performance,and make detailed analyses and also comparisons.Both the derived theoretical and simulation results have verified that the beamforming performance is affected by the covariance of transmitted waveforms for all radar configurations except the PA radar,and the phased-MIMO radar with overlapped subarrays shows the potential of achieving better spatial processing performance compared to its counterparts.Motivated by the supeorities enabled by phased-MIMO radar,we conduct the study on the transmit beamforming technique,namely,transmit beamspace(TB)design,for MIMO radar.We summarize the TB design schemes of MIMO radar into two categories.One is based on designing the covariance matrix of the transmitted waveforms,and the other one is based on devising the TB matrix.Moreover,we provide(also propose)the most representative designing criteria that are of significance for each category,formulate them into mathematical expressions,and then present corresponding solutions and analyses.The essence of the two categories of TB designs has been investigated and revealed.It is worth nothing that we have revised the quiescent TB design method which is based on the radiated energy ratio towards the spatial directions within the sector-of-interest.The modified quiescent TB design leads to lower sidelobes of the generated transmit beampatterns,from which other TB designs based on it can also benefit.Besides,we also propose several other meaningful TB designs,for example,the one based on the consideration of ambiguity function(AF).The MIMO radar utilizing transmit beamforming or TB designs is named by us as the TB-based MIMO radar,and in general,it launches less number of original orthogonal waveforms than the number of transmit antennas and achieves desirable properties via its TB processing.We have defined the AF for the TB-based MIMO radar,and established connections to the well-known Woodward' s AF and the AFs for the standard MIMO radar and the PA radar.This newly defined AF serves as a generalized AF form for which the existing AFs are important special cases.We have also studied the "clear region" that is free of sidelobes for the TB-based MIMO radar AF in order to investigate the resolution performance.Moreover,two limiting cases are identified to bound the "clear region" in the delay-Doppler domain.The actual "clear region" of the TB-based MIMO radar depends on the array geometry and is in between that of the two limiting cases.In addition,we have also compared the TB-based MIMO radar AF with the square-summation-form AF,and have made analyses on the relative sidelobe levels by comparing simulations.We have also studied the problem of clutter and jammer suppression in the contexts of the TB-based MIMO radar and standard MIMO radar configurations.Three jammer suppression schemes that belong to reduced dimension(RD)beamspace processing techniques and two another designs that are based on robust adaptive beamforming techniques are proposed for the suppression under the standard MIMO radar configuration.We employ the MUSIC-based direction-of-arrival estimation to evaluate the performance of the RD beamspace designs,and exploit Lagrange multipliers to derive the power estimates for jammers and targets(desired or interfering ones).Utilizing the spatial differences between the received target and jamming signals and by means of these designs,we show the capability on powerful jammer suppression for MIMO radar.The above-mentioned beamspace designs contain TB designs,and they belong to the regime of spatial processing.For the scenario in which both clutter and jammers are present,we have proposed three-dimensional(3D)space-time adaptive processing(STAP)associated with the TB-based MIMO radar configuraton and the two-dimenstional STAP technique associated with the standard MIMO radar configuration,which respectively deal with the suppression of clutter and jammers simultaneously and in order(i.e.,jammer suppression comes first then slow-time adaptive processing for clutter mitigation).We investigate the rank charactics of clutter and jamming signals for the former joint suppression problem,based on which an efficient 3D STAP algorithm is proposed that significantly reduces the computational burden.For the second joint suppression problem,we derive the correlation function of multipath jamming components after matched filtering at the receiving end of MIMO radar,which can serve as the theoretical measure of matched filtering.Moreover,we propose two minimum variance distortionless response(MVDR)type designs,and provide the closed-form solution for the first MVDR design.